The Mental Lexicon - Volume 12, Issue 1, 2017

Two lexical decision experiments were conducted to study the recognition of two-character Chinese monomorphemic and compound words by adult native Chinese readers. In Experiment 1, the words appeared non-spaced, whereas in Experiment 2 a space was inserted between the two characters. An interaction between word type and spacing reflects a trend for spacing to slow down the recognition of monomorphemic words and speed up that of compound words. The word frequency effect was steeper for monomorphemic than compound words. The number of strokes in the first and the second character influenced the recognition time for compound words, but not for monomorphemic words. The results are interpreted in the light of the parallel dual route model of morphological processing. The holistic route is more prevalent in recognizing Chinese monomorphemic, while the morphological decomposition route is more prevalent in processing Chinese compound words.

The current study examined how non-native speakers process the highly productive derivational morphology of Arabic in which, in contrast to Indo-European languages, word formation involves interleaving a root and template structure. Previous research shows that native speakers of Arabic decompose morphologically complex words in lexical processing. Using cross-modal priming, the current study shows that non-native speakers of Arabic (L1 English) also decompose derived forms such that there is priming between words that share a common root which is not due to semantic or phonological overlap. In spite of the typological distance, native English speakers organize their L2 Arabic lexicons in a manner similar to native Arabic speakers.

Several previous studies have shown that the time-course of word recognition is determined in part by an interaction between connotations of Danger and Usefulness. A small, mostly separate literature has investigated the role of Body-Object Interaction (BOI) in lexical processing. BOI is defined as the ease with which one can interact with an object. To date the lexical decision study of Van Havermaet and Wurm (2014) is the only study to include all three of these constructs. Stimuli in the current study were black-and-white line drawings corresponding to the common nouns used by Van Havermaet and Wurm (2014) . Participants viewed the stimuli one at a time in a random order and had to name them as quickly as possible. Naming times revealed a significant three-way interaction between Danger, Usefulness, and BOI similar to that found for visual lexical decision: The familiar Danger x Usefulness interaction, observed in many previous studies, was observed only for items relatively lower on BOI. The interaction between semantic and embodied processing variables is not restricted to purely linguistic stimuli.

In tasks such as lexical decision, people respond differently to morphologically complex words compared to morphologically simple ones (e.g. in English, lies vs. rise). These divergent responses could conceivably arise from differences in activation levels, or alternatively, from the additional steps required to decompose complex words. To investigate this issue, we used the Deese-Roediger-McDermott (DRM) false memory paradigm, which probes activation of lexical representations by measuring the probability of recalling or recognizing a word (such as lies) after listening to a list of its phonological neighbors (such as wise, lose, lime, etc.). Our results showed a significant false memory effect for complex words, which demonstrates that similar-sounding words can activate representations for stem-plus-affix combinations. Our results also showed no significant difference between false memory rates for complex versus simple words, which suggests that complex stem-plus-affix representations activate at levels equivalent to those of simple stem representations. These findings indicate that differences in activation level probably do not lie at the source of divergent responses to complex and simple words, and that decomposition is the more likely origin.

In some languages the grammatical gender of nouns can be probabilistically detected using formal cues; for instance, in Italian, the majority of feminine nouns end in ‘-a’(e.g., casa, ‘home’) and the majority of masculine nouns end in ‘-o’ (e.g., albero, ‘tree’). It has been hypothesized that the match/mismatch between the formal information of the suffix and the abstract grammatical information on gender affects lexical processing of nouns. An alternative account is that a default option available for gender poses constraints to mechanisms of lexical access for words exhibiting gender markers in the surface form.

In the present study, nouns with highly predictive gender suffix (regular), nouns whose gender cannot be recovered from surface form (opaque) and nouns with misleading gender suffix (irregular) were compared in two reading aloud and two lexical decision experiments. Results confirmed that regular nouns are processed better than irregular nouns. No difference was detected between masculine and feminine opaque nouns.

The results allow the conclusion that a formal gender feature (the gender orthographic regularity) is more likely to affect lexical processing of bare nouns than the activation of a gender default option.

The present study aimed to identify crucial factors that underlie phonological representations in short-term memory (STM) of third-graders with different literacy skills. For this purpose, we used the Nonword Repetition Task (NRT) to disentangle the processing of phonemes’ identity and their serial order. We found no evidence that children’s literacy skills are linked to their capacity for retaining phonemes’ identity. However, their literacy skills are linked to their capacity for retaining phonemes’ serial order. The latter link can be interpreted in terms of a domain-general STM mechanism but is also compatible with the impact of literacy on children’s knowledge of the phonotactic regularities in a language.